The FormalMAGiC workshop invites researchers interested in formalizing computational models of AI or cognitive science to participate in and contribute to the workshop by submitting articles that illustrate their recent research or describe the advances in their work related to the intersection between the two fields.


To design and develop a thinking machine that shows a general level of intelligence, an AGI system should show how to handle (at least a subset of) the most important and challenging (human) cognitive mechanisms and abilities in a more formalized way. This helps the representation, implementation, and integration of a wide range of capacities in computational models of general intelligence. Unless the formal consideration of these capacities attracts the interest of more researchers of the related fields, such as AI and CogSci in particular, what the research community might achieve (e.g. by implementing non-coherent, relatively small fragments of the wide range of the capacities) is a modeling of "narrow-AI" tasks, possibly using newer methods or techniques. In fact, even when achievements of possibly-isolated tasks are reached by means of the most state-of-the-art techniques, it does not mean that the researchers will be able to agree upon working definitions for the mechanisms, nor will they be able to integrate the achievements and conclude common representations or implementation schemes.

Specific mechanisms attracted the attention of scientists in the last decade. Some researchers theoretically investigated the mechanisms; others did experiments and got results about them; other groups have even gone beyond the characterization and started building systems relying on them, or on prototype versions thereof; others proposed ways in which the building may be achieved... etc. However, as far as AGI should be concerned, it is not apparent that a trial to unify such efforts is globally available. Even in the cases when the mechanisms are given particular attention in literature, they do not present a formally coherent topic.

Humans do not only utilize a huge set of uniquely distinguished cognitive mechanisms that affect their intelligence, but they also integrate them in an amazingly smooth manner to accomplish every-day tasks. On the one hand, people may not usually classify what they do as a utilization of their cognitive mechanisms nor an employment of their capacities. They perhaps may not even think about what cognitive mechanisms really are, or what counts as such. Scientific research, on the other hand, should differentiate between the various notions, especially the relatively broad ones, and contribute to giving precision to their vagueness by means of formalization. In this way, modeling, designing, or developing cognitively-inspired AGI systems can be based on a consistent, formal basis.

On the horizon, a point in time is approaching that requires the interested researchers to start focusing on contributions and collaborations that aim at formalizing rigid basics of next-level AGI systems. The motivation to reach this aim is that it was not until scientists continued the development of physical laws of mechanics, for instance, that the building of spaceships or satellites was possible.


The "Formalizing Mechanisms for Artificial General Intelligence and Cognition" (FormalMAGiC) workshop will be held during the Sixth Conference on Artificial General Intelligence (AGI-13), which will co-locate with IJCAI-13 in Beijing, China, between July 31, 2013 and August 3, 2013.

The FormalMAGiC workshop invites researchers interested in formalizing computational models of AI or cognitive science to participate in and contribute to the workshop by submitting articles that illustrate their recent research or describe the advances in their work related to the intersection between the two fields.


The FormalMAGiC workshop discusses the formalization of the most important cognitive mechanisms and human abilities that have already been shown in literature to play essential roles in computational models of artificial (general) intelligence, using available techniques like learning or reasoning.

Cognitive mechanisms and human capacities, in particular those related to general intelligence, have a wide range of possibilities that attract the interest of audience from various (sub)areas in artificial intelligence and cognitive science.

Of special concern to the discussions and presentations during this workshop are the human capacities, such as:
(1) doing abstractions (generalizations or specializations),
(2) creating similarity- and analogy-based decisions,
(3) performing conceptual blending,
(4) creating imaginative images and alternatives to reality,
(5) coherently integrating cognition, or
(6) bridging the gaps between well-known cognitive mechanisms, but the workshop is also open to contributions and treatments of other closely-related capacities.

This workshop seeks initiative contributions that particularly help in demonstrating, formalizing, or implementing the aforementioned capacities (or the related ones) using probabilistic, symbolic, or logic-based approaches. Theoretical characterizations and representation of the essential parts thereof are also very welcome (e.g. characterizing cognitive models and architectures, representing concepts for computing cognition (including quantum structures for cognition), modeling creative capacities in artificial models... etc.). The list is not by any means intended to be exhaustive; the intention is rather to mention the famous mechanisms that already showed importance in the literature of computing cognition, and for which a mature stream of (theoretical and/or practical) research is already established.


1. To point out the most important mechanisms and abilities, that interested researchers believe are needed in achieving a well-established AGI.

2. To discuss, present, and share the possible ways of achieving convergent definitions and formalizations of important cognitive mechanisms among a wide range of related research disciplines.

3. To present and discuss the recent (theoretical or practical) research contributions on the representation, formalization, and/or computation of cognitive mechanisms in AGI systems. In particular, those contributions bridging gaps between several related mechanisms and abilities.


The idea of this workshop is to offer a venue for scientists who are interested in taking the work at the intersection between artificial general intelligence (AGI) and cognitive science (CogSci) to a rigid, more formal level.

Though maybe considered well-characterized by some empiricists, several cognitive mechanisms need to have a formal jargon that allows implementations of AGI systems to base their representation and computation upon (in addition to their possible integration). A convergence to the prospected formal level can be obtained by emphasizing the importance of formally presenting essential cognitive mechanisms and human capacities in developing artificial models of intelligence. These models are expected to implement the mechanisms in order to achieve a human-like level of intelligence.

Important Dates:

  • First CfP:

  • Submission deadline:

  • Acceptance notification:

  • Camera ready version:

  • Workshop: